Hello Daniele, As you know, I am working on support for aggregation on runtime dimension in netcdf. I am now looking into what needs to happen in VariableAdaptor. I am very new at netcdf, so I do not yet understand everything that is happening to the fullest extent in this class.
I have found a way to make it work, basically by ignoring the runtime dimension altogether. Apart from adding it as an ignored dimension, I also had to take into account the possibility of there being four dimensions of which one is the runtime dimension. This is the commit: https://github.com/NielsCharlier/geotools/commit/5278115bd43a77c6509881d59a582e69c8bbe182 The aggregation files now pass and I can publish the layers. But I am still wondering if this is the right approach. Or should perhaps the runtime dimension be treated in the same way as the t- and z-dimensions? Or should there actually be logic in VariableAdaptor for the R-dimension, similar to the logic that exists for the Z- and T-dimensions? What is your opinion? Thanks for your assistance. Kind Regards, Niels ------------------------------------------------------------------------------ _______________________________________________ GeoTools-Devel mailing list GeoTools-Devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geotools-devel
